Figure 7.

a, Representational similarity time course from −200 ms before stimulus to 500 ms after stimulus between each of the map types (saliency and meaning) and the ERP data, but using only the horizontal and vertical EOG channels (after artifact correction). This makes it possible to estimate the effects of any residual eye movement activity. b, Representational similarity time course when the ERP data at each electrode site were computed from the independent components flagged as being generated by ocular artifacts. This makes it possible to assess whether eye movements were systematically related to the physical saliency and meaning maps. a, b, Representational similarity was computed separately for each participant, and the mean across participants is shown here. The upper and lower edges of the gray region represent the upper and lower bound estimates of the estimated noise ceiling (i.e., the highest expected observed correlation).